Monday, January 2, 2012

Is Puijila a pinniped?

Back in February 2009, Natalia Rybczynski and colleagues surprised the paleomammalogy community with their Nature paper naming a new genus and species of early pinniped, Puijila darwini. The holotype skeleton is relatively complete, and include fore- and hind-limbs along with much of the vertebral column, both jaws, and a well preserved skull. Puijila was about one meter long, and would have appeared relatively similar to a modern river otter. It had a short snout and a wide head, with large eyes and relatively high-crowned teeth. The teeth of Puijila still retain many of the cusps lost in modern pinnipeds, and also exhibit pits in the roof of the mouth for the lower teeth (embrasure pits). Unlike modern pinnipeds, it had a long tail, and did not have its fore- and hindlimbs modified into flippers. Puijila was discovered in 2007 from the Haughton Formation on Devon Island in Nunavut (formerly Northwest Territories in Canada). The Haughton Formation was deposited in an impact crater – the impact has been dated to 24-21 Ma (earliest Miocene), and fossil mammals from the Haughton Formation corroborate an early Miocene age. The Haughton Formation was deposited in an ancient lake that filled in the impact crater. In fact – if it were not for the impact, there would be no sedimentary rocks of Miocene age preserved that far North – all the young rocks have been eroded away by glaciation.
The skull and jaw of Puijila darwini, from Rybczynski et al. (2009).

The holotype skeleton of Puijila darwini (from

Previously, the earliest diverging pinniped (and arguably more derived than Puijila) is Enaliarctos, a fur seal sized pinniped from the latest Oligocene and early Miocene of California and Oregon. Enaliarctos retains carnassials, although many of the other dental features are very simplified and reduced, trending toward the condition in modern pinnipeds. Enaliarctos also exhibits limbs developed into flippers – and is very clearly a pinniped. But the relationships of Puijila – to pinnipeds and other carnivores – are not so clear. Because of the older age of Enaliarctos and its marine occurrence, Puijila is hypothesized to represent a lineage of early pinnipeds that stayed in their freshwater niche while marine pinnipeds like Enaliarctos evolved, remaining otterlike. It suggests that pinnipeds went through an otterlike freshwater stage before invading the ocean. Prior to this, Enaliarctos suggested a direct to saltwater invasion – although due to the absence of intermediates, it was not exactly clear one way or the other.

The skeleton and life restoration of Enaliarctos mealsi, from Berta et al. (1989).

Before we continue – I must also be specific about some clade names. Although Rybczynski et al. (2009) refer to Puijila as a member of the Pinnipedia – which is not really the traditional cladistic nomenclature for basal pinnipeds. Annalisa Berta and colleagues (1989) proposed the clade Pinnipedimorpha, for Enaliarctos and all later diverging pinnipeds. Berta (1994) later proposed the name Pinnipediformes for Pteronarctos and all later diverging pinnipeds. Pinnipedia is nested within Pinnipediformes, and Pinnipediformes within the Pinnipedimorpha. Within this traditionally accepted and utilized framework, Puijila’s obviously more primitive morphology than Enaliarctos indicates it should be referred to as a pinnipedimorph.

Rybczynski et al. (2009) listed six characteristics that unite Puijila with Enaliarctos and other pinnipeds. These are: a posteriorly expanded palate (the palate extends posteriorly past the upper toothrow in pinnipeds), an enlarged infraorbital foramen (occurring within pinnipeds due to larger whiskers and greater innervation of the snout), a shelf-like protocone on the upper fourth premolar (occurring in basal pinnipeds and some related arctoids), an upper second molar that is reduced and positioned medially to the upper first molar (reduction of the molariform teeth to conical teeth is a major dental transition within the pinnipedimorpha), a posterodorsally expanded scapula (a feature of pinnipeds, which often have very broad scapulae, an adaptation for swimming), and an expanded deltopectoral crest of the humerus (another feature in pinnipeds related to swimming).

Some of these features may be of only limited use in hypothesizing a close relationship between Puijila and pinnipeds. First, an enlarged infraorbital foramen (the small hole below the eye socket in a skull) also occurs in many mustelids, such as badgers (Taxidea), weasels (Mustela), and most (if not all) otters (e.g. Lontra, Enhydra); in badgers and weasels, this is due to a more innervated and sensitive snout, an adaptation for digging in burrows. This characteristic may not be useful in identifying early pinniped relatives, as otters (another hypothesized pinniped sister taxon) exhibit this feature – presumably evolving for the same purpose. The two postcranial features – an enlarged deltopectoral crest and a posterodorsally expanded scapula – may not be coded correctly in Puijila. To be quite honest, based on the available photographs – these features do not appear too different between Puijila and a river otter (Lontra), and don’t really exhibit the derived condition, which is much more extremely developed in Enaliarctos and other fossil and modern pinnipeds). This leaves only three reliable characters behind.

Skeletons of Enaliarctos (A), Puijila (C), and Lontra (D).

One of these features – a posteriorly expanded hard palate – also appears in the aberrant arctoid Kolponomos, as well as Pachycynodon and Allocyon (Tedford et al., 1994). This feature appears to be more widely distributed than previously admitted, although it is worth mentioning that it is probably an adaptation towards aquatic feeding. The two dental characteristics are probably more reliable, although it is odd to note that these two dental characteristics are also found in Kolponomos. Kolponomos will feature more in part two of this post, as it is both a fascinating creature which I have not yet properly covered on this blog, and it is also extremely topical to the subject of Puijila.

A wonderful reconstruction of Kolponomos by Ken Kirkland, published in Neptune's Ark by David Rains Wallance.

There are a whole host of pinniped characteristics that Puijila does not have, or features that were not described in enough detail to independently evaluate. Cranial and dental features that characterize pinnipeds (either at the level of the Pinnipedimorpha, Pinnipediformes, or Pinnipedia) include a reduced upper first molar, reduced cingulum on the upper first molar, an absent or reduced metaconid, loss of the embrasure pit between the upper fourth premolar and first molar, closely spaced mastoid and paroccipital processes, a reduced nasolabialis fossa, reduced cingulum on the upper third incisor, and a reduced trigonid cusp on the lower molars. None of these derived conditions appear to be present in Puijila. A number of postcranial features which characterize pinnipeds (either at the level of the Pinnipedimorpha, Pinnipediformes, or Pinnipedia, as above) include a short and robust humerus, enlarged tuberosities of the humerus, a flattened radius and ulna that are longer than the humerus, emphasis of certain digits in the hand and foot (i.e. largest digit in the hand being the first or “thumb”, and the lateral and medialmost digits of the foot being largest and the smallest being the middle digit), a short, flattened femur with an enlarged greater trochanter and medially inclined condyles, a reduced tail, and a short ilium (bone in the pelvis). Curiously, none of these features are present in Puijila – although nearly all of them (perhaps all – I can’t remember off the top of my head) occur in Enaliarctos.

In summary, there are very few features that actually unite Puijila and pinnipeds. The cladistic analysis of Rybczynski et al. (2009) was relatively limited, both in terms of the number of fossil and modern species used, and also with regard to the number of morphological characters used in the analysis. An earlier study – considered to be the landmark phylogenetic analysis of pinnipeds, Berta and Wyss (1994) – utilized 143 characters, while Rybczynski et al. (2009) only used 42. Granted, the earlier study focused on relationships within pinnipeds – and many of those characters may or may not apply to Puijila and other similar arctoids.

Because of the above, I am skeptical that Puijila has a close relationship with pinnipeds – and although I’m not convinced, I am equally receptive to the idea pending further analyses. There is much we’ve been waiting for – a detailed description of its anatomy and large, detailed figures of its skeleton were not included in the original study, which is understandable given the short length of Nature papers (however, it was not included as supplementary info either). There is certainly more work to be done, and there are rumors of a more detailed phylogenetic analysis in the works; such a study should include a comprehensive description of the skull and skeleton of Puijila accompanied by exhaustive figures (formerly difficult, but several modern journals have made this a far less difficult prospect), sample a broader variety of caniform carnivorans as well as more fossil and modern pinnipeds (fossil pinnipeds with postcranial skeletons such as Allodesmus, Thalassoleon, Leptophoca, Acrophoca, Piscophoca, Imagotaria, Neotherium, and Proneotherium) as well as the “beach bear” Kolponomos (which was not included in the original analysis or discussed by Rybczynski et al. 2009). Puijila certainly could belong to some group of otterlike “protopinnipeds” as hypothesized, but given the few characters uniting them, it very well could be some sort of mustelid or other aquatic carnivore.

Further reading

Official Puijila website (Canadian Museum of Nature)

Puijila, a very basal 'pinnipedimorph' (this blog)

Commentary by Ed Yong (Not exactly Rocket Science)

Commentary by Brian Switek (Laelaps)

Literature cited

Berta, A., C.E. Ray and A.R. Wyss. 1989. Skeleton of the oldest known pinniped, Enaliarctos mealsi. Science 244:60-62

Berta, A. 1991. New Enaliarctos* (Pinnipedimorpha) from the Miocene of Oregon and the role of "Enaliarctids" in Pinniped Phylogeny. Smithsonian Contributions to Paleobiology 69.

A. Berta. 1994. New specimens of the pinnipediform Pteronarctos from the Miocene of Oregon. Smithsonian Contributions to Paleobiology 78:1-30

Rybczynski, N., Dawson, M.R., Tedford, R.H. (2009). "A semi-aquatic Arctic mammalian carnivore from the Miocene epoch and origin of Pinnipedia". Nature 458 (7241): 1021–24

R. H. Tedford, L. G. Barnes, and C. E. Ray (1994). "The early Miocene littoral ursoid carnivoran Kolponomos: Systematics and mode of life". Proceedings of the San Diego Society of Natural History 29:11-32.


Lee said...

Thoroughly argued and well written. Waiting for pt. 2.

J. Velez-Juarbe said...

Nice take on this critter! I am too awaiting for a longer description and its inclusion in a more comprehensive phylogenetic analysis. I'm skeptical of its pinnipedimorph affinities and cannot help myself thinking about otters every time I see Puijila...

Robert Boessenecker said...

Indeed! I've heard that this study is in the works, but I'm not sure when we'll see it. We may be waiting a while.

Furthermore - Puijila does indeed remind one of a lutrine otter. In the original paper, Puijila is stated to lack the derived features of lutrines, and does not plot with them on the cladogram. However, I don't really know lutrine morphology terribly well, and I don't quite have the time to go character by character through the matrix and sort them out, so I'm not quite sure what lutrine synapomorphies it lacks.

J. Velez-Juarbe said...

That's the one thing that bothers me when people describe new important species/specimen in Nature or Science, the provided space is too small, and I'm not really fond of dumping everything else in the supplementary online material files. Then detailed descriptions take a long time... I know of a particular example that's been 11 years in the making...

Raymond Pierotti said...

Pujila makes much more sense as a pinniped if you consider it to be an ancestral phocid rather than an ancestral pinniped. This of course requires rethinking pinniped evolution and reconsidering the original diphylyetic origin proposed by McLaren (1960, 1975) rather than the monophyletic sheme proposed by Berta and Wyss. Tghis leaves Enaliarctos where it was all along, as an ancestral Otarioid.

One thing none of the monophyletic folks have never explained is why there are no Miocene Phocid fossils in the Pacific. I know they have tried to force fit the Desmatophocinae into this role, but these do not fit well at all. This has been proved rather convincingly by recent work by Koretsky et al. 2016.

Robert Boessenecker said...

Not really, in my opinion - it doesn't have a single phocid synapomorphy (thickened bulla, V-shaped frontonasal suture, widely flaring and everted ilium, among others), and is arguably more distantly related to phocids than Enaliarctos is.

Not a single cladistic analysis has ever recovered Enaliarctos within Otarioidea, and not a single cladistic analysis has ever demonstrated pinniped diphyly - apologies to the diphyly crowd, but Koretsky et al. (2016) has a lot of problems and is ultimately not a convincing or effective body of research.

Raymond Pierotti said...

So you don't know lutrine morphology and you are willing to argue that Puijila is an otter.

Like most monophiles you are ignoring Biogeography, which cannot be ignored. Where are the early phocids from the Pacific. Enaliarctos also lacks phocid synapomorphies, so where did these traits come from. For Alldesmus or other Desmatophocines to be related to phocids they would have to reverse their mode of locomotion, both on land and in water.

If you do a an analysis based on ecology, behavior, mating system, physiology, and developmental patterns, not to mention biogeography, diphyly seems pretty obvious. What are the "problems" with Koretsky et al (2016) to which you allude.

Robert Boessenecker said...

Oh dear, there's a lot to unpack here.

1) I never said Puijila was a lutrine. I said it was "otter like" in its skeletal anatomy, which is objectively true, and made comparisons with lutrines. Please do not put words in my mouth.

2) Of course biogeography should not be ignored - but it should not inform the topology of the evolutionary tree. Rather, the tree and the tips of the branches tell us about the biogeographic history within a phylogenetic framework.

3) Allodesmus is the one forelimb-rowing phocoid pinniped - and it is highly derived. Enaliarctos and Odobenus have adaptations for an intermediate form of locomotion using both fore- and hind-limb dominated swimming. If the common ancestor of desmatophocids and phocids was similar to Enaliarctos, it probably also had the primitive locomotor pattern.

4) On the topic of locomotion - you can't just overemphasize the features you like or think are informative, because then the phylogenetic results become skewed by our own biases. That's why I always aim to include the most character data possible. And the result of a modern approach to pinniped cladistics is a desmatophocid + phocid clade (Phocoidea). That being said, I would not be surprised if desmatophocids were otarioids - but I won't believe it until it shows up in more phylogenies. It hasn't shown up in mine, so I don't really believe in the concept.

5) This is exactly the problem with Koretsky et al. (2016) - the character matrix is cherry picked to only a few taxa and a handful of characters the authors know will force a diphyletic topology. This is highly misleading.

6) I'm not sure what your point is about Enaliarctos not having phocid synapomorphies. It's never been considered to be a phocid by anyone, so it comes as no surprise that it doesn't have any. It doesn't have any otariid or odobenid synapomorphies either - probably because it is ancestral and pretty far down the tree.

7) The "total evidence" analysis you speak of would almost certainly support pinniped monophyly. Ecology: pierce feeding fish eaters, marine (unlike all other carnivorans). Behavior/mating system: polygynous and sexually dimorphic (reversed in phocines), to an extreme degree not seen in any terrestrial carnivorans. Developmental patterns: long single fetus gestation, milk teeth reduced and shed early.

8) DNA. All molecular analyses unanimously support pinniped monophyly, to the point where I don't believe having this conversation is even a good use of my spare time.